INDEX

Explanations

dates

oai_token-act-pair · gpt-3.5-turbo

instances of a particular word sequence or branding reference

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

bin

-0.70

 Proceedings

-0.67

 Tactics

-0.66

doi

-0.65

 IEEE

-0.64

 appar

-0.64

 Transcript

-0.64

framework

-0.63

 Init

-0.62

False

-0.62

POSITIVE LOGITS

irst

3.41

elve

2.16

DEA

1.15

CAR

1.13

DER

1.03

 Hend

0.98

 mustard

0.98

irteen

0.93

essors

0.92

etermined

0.89

Activations Density 0.033%

dates

instances of a particular word sequence or branding reference

No Comments

No Known Activations